Cytomegalovirus (CMV) retinitis is the most common ocular opportunistic infection observed in patients with AIDS, and development of CMV retinitis depends on immunodeficiency. In normal, non-immunosuppressed individuals, CMV is usually latent, and if the virus does reactivate, presumably the virus-infected cells are quickly eliminated by the host's immune system. In contrast, in the immunosuppressed host, such as patients with AIDS or those patients who are systemically immunocompromised following an organ transplant, a bone marrow transplant, or during chemotherapy, CMV can cause disseminated disease with both systemic and ocular manifestations. The inability of the host to mount an effective immune response to the virus likely contributes much to the pathogenesis of this disease by allowing the virus to replicate, and both systemic and local infections with CMV are associated with immunosuppression. During the first funding period, we demonstrated that inoculation of 5 x 10/2 PFU of MCMV into immunosuppressed BALB/c mice (steroid-treated or T cell depleted) resulted in fulminant retinal necrosis in the majority (>90%) of the mice. Virus became latent at several sites, including the eye, and could be reactivated by immunosuppression. During latency, a majority of the latently-infected eyes were positive for transcripts of the MCMV immediate early gene 1 (IE1). The murine model of MCMV retinitis will be used in the studies proposed in this application to: (1) determine whether immunosuppression-induced reactivation of latent MCMV in the eye results from in situ reactivation of ocular virus, from hematogenous spread of reactivated virus from extraocular sites, or from a combination of in situ reactivation and hematogenous dissemination of reactivated virus, (2) determine whether production of IE1 mRNA is a marker for MCMV latency in the eye, and (3) determine whether adoptive transfer of NK cells, macrophages, and/or virus-specific T cells protects against MCMV retinitis following supraciliary inoculation. Results from studies to identify sites of MCMV latency and to use a cellular approach to prevent MCMV retinitis may help us understand the pathogenic mechanisms of CMV retinitis in human patients. Results of experiments to see whether IE1 transcripts are markers of MCMV latency in the mouse eye may ultimately be useful for predicting which CMV-seropositive, HIV-1 infected patients are at the highest risk of developing CMV retinitis.